Interesting. Implemented the Index of Coincidence version for a class in college. Was fun.<p>This seems like the kind of problem that could be better solved by coming up with a better algorithm. Like, ultimately the model is performing an "algorithm" of sorts on the data to guess the key length.
Inverse:
“CryptGPT: Privacy-Preserving Language Models Using Vigenere Cipher”<p><a href="https://huggingface.co/blog/diwank/cryptgpt-part1" rel="nofollow">https://huggingface.co/blog/diwank/cryptgpt-part1</a>